Adrien Danel

Study of the Volatile Organic Contaminants Absorption and their Reversible Outgassing by FOUPs

Today, the use of Pods or FOUPs (Front Opening Universal Pod) in IC manufacturing leads to specific molecular contamination issues related to the enclosed environment made with porous polymers (mainly PEEK, PC and PP) that constitute these containers. Indeed, such materials are known to outgass airborne molecular contaminants (AMC), especially polymers additives [1,2]. They are also able to absorb volatile compounds present in their atmosphere coming from the connection to an equipment or from the release of wafers just processed [3,4]. As a result, a reversible outgassing of species previously trapped in plastic is possible. This is especially critical in presence of wafers sensitive to the released contaminants leading then to potential detrimental impacts. This cross-contamination scheme was clearly evidenced for volatile acids in presence of Cu layers leading to corrosion issues [4].

Mapping of Metallic Contamination Using TXRF

Two main routes drive the needs for metallic contamination analysis in Integrated Circuit (IC) manufacturing: the insurance of high yields and the introduction of new materials used to target specific electrical, optical or mechanical properties in advanced microelectronic, including non volatile memories and above IC features. With metallic contamination specified by the ITRS at 5E9 at/cm 2 for critical metals at critical production steps of nodes 90nm and beyond, analytical methods should offer capabilities in the E8 at/cm 2 range, this for a set of elements with Z ranging from Na to Pb.

Influence of Cobalt Contamination in the Measurement of Diffusion Length of Silicon Wafers

Si bulk lifetime techniques are really sensitive to Co after dr ive-in step performed by RTP: the bulk diffusion length is strongly reduced, even at very low level of contamination. In case of p-type silicon, we point out that low energy photodissociation enhances this decrease, showing a pairing behavior of Co that has demonstrated to be with Boron, as for t he well known Fe-B pairs. A detailed study of the interaction of Co with Si which has been foc used on the detection of Co in Si using bulk lifetime methods, especially the SPV technique has been ca rried out. Evidence of recombination centers related to Co by Deep Level Transient Spect roscopy (DLTS) is shown, along with the activation of Co, its recovery kinetics in p-type Si and the low limi t of detection of the SPV technique.

HF Last Passivation for High Efficiency a-Si/c-Si Heterojunction Solar Cells

Amorphous/crystalline silicon heterojunction solar cells are commonly made by low temperature deposition of front and back side thin films on bare H-passivated Si wafers, obtained by HF last processes. This work discusses the impact of HF last step parameters on cell performance, considering textured and cleaned Si (100) wafers. A complete native oxide removal is mandatory and achieved in a short time (< 5 min) by HF concentration higher than 1% (by weight). Above 1%, surface passivation and cells performance slightly increases with the concentration. The best process time is found to be the minimum time to deoxidize textured wafers, as seen by a good dewetting. For [H > 2% this is less than 1 min. Longer process times slightly degrade surface passivation. Post rinse and drying, provided they do not reoxydize the surface, were seen to have no impact. The delay between the HF last and deposition steps is critical and depends on the efficiency of the cleaning before the HF last. With a high performance cleaning, leading to a very good surface passivation (< 10 cm/s surface recombination velocity), 30 min delay has no impact and 90 min leads to about 5% relative degradation of cell performance. Regarding the HF cleanliness, HCl spiking is an efficient way to enhance robustness of surface passivation keeping < 10 cm/s values when the metallic contamination, including Cu, is in the sub 50 ppb range.

How to Overcome the Effects of Silicon Build-Up during Solar Cell Wet Chemical Processing

Although the chemical reaction is well known, the anisotropic etching of Si in alkaline solutions is a complex process. This is particularly true in the solar industry where a large mass of silicon is typically introduced into the etch bath. The etch by-products (silicates) affect the balance of the etching specie. If adequate compensation is not made for these by-products, a significant drop in etch rate and an increase in contamination levels is typically noticed. Because of this contamination, production lines would suffer from unpredictable wafer characteristics and hence lower cell performance.

Estimation of Detrimental Impact of New Metal Candidates in Advanced Microelectronics

The increasing complexity and miniaturization of integrated circuits (IC) requires the introduction of a large number of new materials which represent possible risk of contamination. Indeed many integration steps use “exotic metals” to achieve the targeted device performance: - The transistor includes high-k dielectrics to replace SiO2, silicides to replace the polycrystalline Si gate, metals for electrodes, substrates with high mobility. - Interconnects need new barriers for Copper. - Non-volatile memories and Above IC components such as RF features or imagers introduce new materials to target specific electrical, magnetic or optical properties. This study follows work published in 2005 by Bigot et al. [1] and will focus on the behavior of each element toward Si and SiO2 properties. This work aims at completing previous work dealing with the seriousness of cross contamination and will report for the first time with “news” metals for which the behavior in Silicon and oxide is really unknown ( Sc, Er, Yb, La, Cd,...).

Deposition Behavior of Volatile Acidic Contaminants on Metallic Interconnect Surfaces

Introduction With the decrease of semiconductor device geometries, airborne molecular contaminants (AMC) deposited on wafer surfaces have become an increasing concern [1]. Among AMC, acidic compounds are well known as source of particle formation (time dependent haze) and corrosion of metal wiring [2,3] leading to potential detrimental effects on devices. However, very few data are available on their deposition behavior [2]. In this work, we have investigated the deposition behavior of volatile acids on metallic interconnect surfaces in order to define their affinity and to get a better evaluation of their potential detrimental impact. More precisely, for the first time to our knowledge, the sticking of the hydrogen fluoride (HF), the hydrogen chloride (HCl) and the hydrogen bromide (HBr) have been studied on general interconnect materials: aluminum and copper.

Surface Potential Difference Imaging Applied to Wet Clean Monitoring

The monitoring and optimization of wet clean and surface preparation processes is a major challenge in the microelectronics industry [1, 2]. Today, the main methods used in clean rooms are visual inspection by light scattering (principally applied to particle detection) and metallic contamination detection by Total-reflection X-Ray Fluorescence (TXRF). These methods, despite good sensitivity and recent progress [3, 4] are not sufficient, especially considering non-visual defects not measurable by light scattering, nor TXRF due to their chemical nature or to their size and location (TXRF is not applicable to light elements – with Z < 11 – and is typically a 1 cm resolution tool, with 1 to 2 cm edge exclusion). Non-vibrating Surface Potential Difference Imaging (SPDI), introduced in 2005 under the name of ChemetriQ® is an in-line, non-contact, non-destructive inspection technique based on the imaging of surface Work Function (WF) lateral non-uniformities [5]. Recent studies show very promising results for SPDI: high sensitivity to traces of metals on Si wafers with native oxide [6]; fast imaging capabilities of unpatterned or patterned wafers with sensitivity to chemical residues and charge [7, 8]. In this work, the ChemetriQ method is evaluated for in-line control of wet clean processes. The variation of SPDI data from various contaminants is compared to intra- and inter-wafer variations related to the cleaning and measurement conditions. Note that all wafer maps are presented with the notch oriented at 6:00.

Impact of Metallic Contamination on Si: Shortloop Issues

Context and Motivations The increasing complexity and miniaturization of integrated circuits (IC) requires the introduction of a large number of new materials which represent potential risk of contamination. Indeed many integration steps use “exotic metals” (high K, silicides...) to achieve the targeted device performances. Thus, considering also the “conventional” contaminants related to equipments, fluids and human activity, around 50 metallic elements might have to be considered in advanced microelectronics. The works published in 2005 by C.Bigot et al. [1], and in 2006 by Y.Borde et al. [2] deal with, on one hand, the classification of each metallic element according to their properties toward Si and SiO2, and on other hand, their dangerousness. However, the classification and the dangerousness estimation strongly depend on the shortloop conditions. This paper investigates how the major shortloop issues impact metallic contamination studies.

Copper decontamination ability of supercritical-CO2/additives on CVD and spin-on porous MSQ materials

The paper focused on the development of very dilute mixtures of chelating agents, acids and organic solvents for post etch residue (PER) removal and copper (Cu) low-k decontamination under supercritical CO 2 (SCCO 2 ) foradvanced nodes (< 65nm) BEOL integration. The Cu low-k decontamination ability of each mixture was carried out on Spin-On Dielectric (SOD) and Chemical Vapor Deposition (CVD) porous low-k. The copper decontamination ability of SCCO 2 /additives systems were also studied on ashed and unashed low-k blanket wafers. Finally, the paper presented the Cu decontamination performance and Cu PER removal ability of SCCO 2 /additive systems compared to conventional wet chemistries.